I'm trying to power 7-12 servos, and I was under the impression that each one would need about an amp, but in looking around for an appropriate BEC to supply them, I notice that most seem to output around 1-3.5 amps.

They won't all be running at once, but often, say 4 will be drawing enough juice to move.

Obviously, I'm missing some link in my understanding. How do I determine how many amps will be needed from the power supply?


Forgive me I seem to get caught up in people looking for answers to a question they didn't ask a few times already. So given this question;

"How do I determine how many amps will be needed from the power supply?"

You can roughly guess it, in your case you say 4 servos drawing an amp each, 4 amps total.

Pick any BEC that will supply greater than 4 amps, there are plenty available from the other answer I already gave.

Then apply a logging circuit to validate your estimate; Make your own that logs both current and voltage, or buy a simple already build device that is made just for that purpose such as an Eagle Tree Logger

There is no magic formula to find correct amperage. Only real world usage and observation will give you a more precise estimate. Any such formula would not account for your mechanical design, friction, drag, gravity etc.

  • $\begingroup$ The data sheet I have doesn't actually specify how many amps these servos want. I guess I was hoping that for a standard size servo, 1A was a reasonable estimate for what it wants. One of the BECs I was looking at has a 3.5A output, but claims it can supply 8 servos. $\endgroup$ – Khrob Jan 9 '13 at 6:25

There's two possible current requirements to consider here.

Ideally, you need to be able to supply enough current to account for the rated stall current (that is, the current drawn by the servo at its maximum rated torque) of all of your servos simultaneously. Imagine a bunch of your mechanical components getting jammed on each other in some gruesome fashion and busting your servos. You don't want to add insult to injury and fry your BEC in the process as well. If you don't have that number from the datasheet for the servo, you can test it by stalling a servo manually. This number can be quite a bit larger than your normal operating current, often to the point of being impractically large to design for. In the (unlikely) event that you can somehow prove that given servos in your application will never be placed in a stall condition, you can skimp on that a bit. Keep in mind that a servo doesn't have to be actively moving on its own to be stalled; a force trying to move it from its set position can do it as well.

The second, softer requirement is your real-world operating current, which is a lot more application-specific. The amount of current drawn by a motor is dependent on the amount of torque it is exerting to move the load. Measure the current drawn by each servo while moving its real-world load. Add those up for any given group of servos that might be moving at the same time. The largest of those is your maximum normal operating current. Multiply that by 1.5-2 for a safety margin.

See also: What is the best way to power a large number (27) servos at 5 V?


I'm not sure why you are trying to make it hard than it is. If you have 4 servos that are in use at one time, that draw an amp each, you need 4 amps, period. BECs are convenience components, that are used because R/C servos used to burn out if provided with more than 5 volts. 6 Nicad batteries, which was the most common configuration before LiPo technology came around, worked out to 9.6 volts (6 X 1.2v). BECs are available in many sizes, and greater than 3.5 amps is common.

If the requirement is greater than what is available, you simply provide your own voltage/amperage to the servos - often straight from a 6v battery, again the BEC is/was just a convenience component. If you do this with excessive current (more than the controller can pass through) you need to remove the red wire from the plug that goes into the controller, and supply voltage via that red wire. the black wire, needs to be both a ground connection for the signal, so it can not be removed from the controller, but it also needs to be the negative connection for power. The need to do this is pretty uncommon. The controller power connections are just straight pass through.

If you are using a plain R/C receiver the rule of thumb is the circuit board can handle about 6 amps max. A good digital servo will draw 2 amps at full power, but in most cases, what I have seen used in robotics are not good servos, so your guess of about an amp is reasonable.

  • $\begingroup$ Interesting background, but a given servo model DOES have a required operating voltage that his battery might fall outside of. Given that he's researching BECs to power the servos, I would imagine he has some higher voltage elements to his design operating off a battery directly (say, a 12V motor). $\endgroup$ – Joe Baker Jan 5 '13 at 8:03
  • $\begingroup$ He may have a servo whose voltage limit is outside of his supply - how does my answer in any way conflict with that? BTW every servo I have ever seen has a voltage limit, but not a specific single voltage requirement. More Volts results in faster servo. That has been my experience from \$2 R/C junk to professional \$5000 CNC servos. $\endgroup$ – Spiked3 Jan 5 '13 at 11:51
  • $\begingroup$ Just that it's not a "convenience component" if you have a servo that doesn't take your battery voltage. There's plenty of hobby servos with upper limits around 6V. $\endgroup$ – Joe Baker Jan 5 '13 at 13:50
  • $\begingroup$ It is the convenience of not using a separate power source, which is what was done before BECs came along. BECs are not a needed component, they were originally intended to save weight from a second battery source. As far as I know, that is still their only purpose. $\endgroup$ – Spiked3 Jan 5 '13 at 18:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.